Re: linux-next: build failure after merge of the rcu tree

From: Paul E. McKenney
Date: Mon May 22 2023 - 10:57:07 EST


On Mon, May 22, 2023 at 11:45:03AM +1000, Stephen Rothwell wrote:
> Hi Paul,
>
> On Thu, 18 May 2023 19:12:52 -0700 "Paul E. McKenney" <paulmck@xxxxxxxxxx> wrote:
> >
> > On Fri, May 19, 2023 at 10:59:50AM +1000, Stephen Rothwell wrote:
> > > Hi all,
> > >
> > > After merging the rcu tree, today's linux-next build (x86_64 allmodconfig)
> > > failed like this:
> > >
> > > kernel/rcu/rcuscale.c:340:27: error: 'get_rcu_tasks_trace_gp_kthread' undeclared here (not in a function); did you mean 'show_rcu_tasks_trace_gp_kthread'?
> > > 340 | .rso_gp_kthread = get_rcu_tasks_trace_gp_kthread,
> > > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > > | show_rcu_tasks_trace_gp_kthread
> > >
> > > Caused by commit
> > >
> > > 9bb839a83e1b ("rcuscale: Measure grace-period kthread CPU time")
> > >
> > > CONFIG_RCU_SCALE_TEST=m
> > >
> > > I have used the rcu tree from next-20230518 for today.
> >
> > Huh. Modules and #ifdefs trip me up again. Will fix, and thank you
> > for catching it!
>
> This is now:
>
> ERROR: modpost: "get_rcu_tasks_trace_gp_kthread" [kernel/rcu/rcuscale.ko] undefined!
>
> I have used the rcu tree from next-20230518 again for today.

-ENOREPRODUCE

Perhaps the #ifdef around its definition is causing trouble in your
environment. And to be fair, the small-memory systems that might not
want that function unconditionally defined could do even better by not
configuring anything requiring Tasks Trace RCU.

I will adjust. Here is hoping!

Thanx, Paul